Plot Lookbook

In the realm of sleep studies and college students, there is a prevailing narrative that suggests a direct correlation between sleep duration, sleep quality, and academic performance/cognitive function. A specific study conducted in 2019 with MIT freshman tracked and analyzed their sleep and academic performance throughout a semester to understand and identify the relationship between these variables.

In the MIT study, the measured variables included gender, sleep duration in minutes, mean sleep quality (a score from 1-10), standard deviation of sleep minutes, standard deviation of sleep quality, and the overall score (sum of all grade relevant quizzes and exams). In my own study, I chose to measure the same variables, along with a couple additional ones, to identify whether those factors had any relationship with test score/academic performance. The variables I looked at were: Sleep Duration (minutes), Bed Time, Wake Up time, Sleep Quality, Heart Rate Variability (HRV), Exercise Minutes, Steps, and Overall Score (average of all grade relevant exams).

Upon closer examination of my own data, I noticed that there were many nuances that told a different personal story than the findings from MIT, which will be shown throughout this analysis.

I initially chose to examine the distribution of my sleep quality data and MIT’s sleep quality data, to identify similarities and differences between the two. Sleep quality is considered one of the more influential factors of academic performance. Upon looking at the plots, my data showed a larger range of sleep quality and a lower median sleep quality score when compared to the MIT dataset. The variability in my dataset challenges the generalizations from broader studies like the one conducted by MIT, emphasizing the need for personalized insights and the unique individualities of each person.

After comparing both sleep qualities, I decided to focus on MIT’s dataset, to identify whether there was a difference in sleep quality between males and females. In looking at the plot, it is clear that female’s have a larger variation in sleep quality, which is consistent with my own distribution from above. While females seem to have a higher median sleep quality score, they also are more susceptible to lower sleep quality score, consistent with research conducted by others on females’ vulnerability to poor sleep quality. To investigate this further, I decided to look at sleep duration between females and males.

In this interactive scatterplot (link), users can look at the relationship between sleep duration in minutes to overall test scores in the MIT dataset, making comparisons between females and males. Upon closer inspection, it appears as if sleep duration may be more of a predicitve factor in overall score and academic performance, than the sleep quality of genders. While, overall, females had a lower sleep quality, according to the sleep duration scatterplot, females tended to sleep longer than males, and had higher test scores, which likely means better academic performance overall.

https://rhiannons-128.shinyapps.io/FinalProject_Shiny/

After examining MIT’s dataset, specifically at sleep quality and duration, I decided to compare my own data to see if there was a relationship between sleep quality and duration. As depicted in the line plot, it seems as if there is a direct relationship between sleep quality and duration, as the highs and lows of both plots coincide with each other. This makes sense, as sleep score is calculated based on “how long you slept, how well you slept and evidence of recovery activity occurring in your automatic nervous system derived from heart rate variability.” There was a large range in my own sleep duration over the data collection period, mirroring the variability in the MIT dataset. This illustrates that college sleep schedules can be hectic and are ever-changing depending on the demands of that week.

Diving further into sleep quality, I aimed to examine whether Heart Rate Variability and other factors were closely related to each other and how they influenced overall score. However, upon closer inspection of my radar chart, it is hard to tell with certainty whether HRV has a clear relationship to overall score, as no categories are significantly more pronounced than others. Because of these relatively inconclusive results, I decided to look at whether exercise minutes were related to sleep quality, which in turn could be related to overall score.

In this animated plot, the point size is relative to sleep duration, observing the relationship between sleep quality and exercise minutes throughout the data collection period. After examining the plot’s progression, it does seem that there is a relationship between sleep quality and exercise minutes. As exercise minutes increase, sleep quality also tends to improve, which is supported by numerous studies indicating that “people who engage in at least 30 minutes of moderate aerobic exercise may see a difference in sleep quality that same night.” If this could be further studied in the context of college and academic performance, it could be a compelling topic for discussion concerning wellbeing of students and required courses.

To comprehensively understand the relationship between various variables and overall academic performance, I decided to create a correlation heatmap for both my dataset and MIT’s dataset. As seen in MIT’s dataset, the highest correlation (0.73) is observed between mean sleep quality and sleep duration, aligning with previously mentioned calculation of sleep quality. There is also a positive, although weak, correlation between sleep quality and overall score, indicating that sleep may not be as important as originally assumed in academic performance.

Examining my own data, the variables with the highest positive correlation were sleep quality and sleep duration (0.82), echoing MIT’s dataset, further emphasizing the connection of these two variables. However, the other variables do not seem to follow the same trends that MIT’s dataset does, showing negative correlations for almost all other variables. Looking at overall score, the variable most postively correlated is Heart Rate Variability (HRV), but with an extremely low and practically negligible correlation of 0.07.This highlights the limitations in correlations with my data, making it challenging to draw definitive conclusions as there were in MIT’s study.

MIT concluded that “longer sleep duration, better sleep quality, and greater sleep consistency were associated with better academic performance.” While I would like to say the same about myself, it is challenging to say so with certainty due to a lack of concrete evidence as demonstrated in my plots. While I could make generalizations as I already have, my data emphasizes the point that sleep and academic performance are more personalized than the generalized statements found not only in MIT’s study, but in other studies that include college students.

The intricacies of sleep patterns, exercise routines, and academic performance are highly individual, suggesting that a one-size fits all approach may not be the best. Everyone lives, studies, and performs differently in various academic situations. If one wanted to understand how their own sleep affects their academic performance, I would recommend conducting their own study over an entire semester to obtain the most comprehensive results.

After completing my study and analysis, I have concluded that my own sleep doesn’t affect my academic performance as much as it affected the MIT freshman. Whether this is because of my own lifestyle or other factors is not entirely known; however, recognizing the importance of sleep, I will strive to maintain a healthy sleep schedule in order to be at my best day to day.